API Overview

Archyl provides a comprehensive API that allows you to integrate architecture documentation into your workflows, tools, and automation pipelines.

API Endpoints

Archyl offers two API interfaces:

REST API

The REST API provides full access to all Archyl features:

  • Create and manage projects
  • Add, update, and delete architecture elements
  • Manage relationships
  • Handle ADRs and documentation
  • Export diagrams
  • Run, steer, and schedule managed agents

Base URL: https://api.archyl.com/api/v1

MCP Server

The Model Context Protocol (MCP) server enables AI assistants to interact with your architecture:

  • Claude Code, Claude Desktop
  • Cursor
  • VS Code with Copilot
  • Other MCP-compatible tools

HTTP Endpoint: https://api.archyl.com/mcp

Authentication

All API requests require authentication using an API key:

curl -H "X-API-Key: your-api-key" \
  https://api.archyl.com/api/v1/projects

Creating API Keys

  1. Go to your Profile → API Keys
  2. Click "Create API Key"
  3. Choose permissions (read-only or read-write)
  4. Copy and securely store your key

Key Permissions

Permission Description
Read View projects, elements, and documentation
Write Create and modify projects, elements, relationships

Quick Start

List Your Projects

curl -X GET \
  -H "X-API-Key: your-api-key" \
  https://api.archyl.com/api/v1/projects

Create a System

curl -X POST \
  -H "X-API-Key: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "E-commerce Platform",
    "description": "Main e-commerce system",
    "type": "internal"
  }' \
  https://api.archyl.com/api/v1/projects/{projectId}/systems

Create a Relationship

curl -X POST \
  -H "X-API-Key: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "sourceId": "system-1",
    "targetId": "system-2",
    "label": "Sends orders to",
    "technology": "REST/HTTPS"
  }' \
  https://api.archyl.com/api/v1/projects/{projectId}/relationships

Managed Agents

These endpoints let you start, follow, and schedule managed agent runs from your own tools. Paths are relative to the base URL.

Profiles and Skills

Method Path Description
GET /agents/skills List the built-in skills a profile can enable
GET /agents/profiles List agent profiles (a default profile is created if there is none)
POST /agents/profiles Create a profile
PUT /agents/profiles/{id} Update a profile
DELETE /agents/profiles/{id} Delete a profile and pause the schedules that use it

Runs

Method Path Description
POST /projects/{projectId}/agents/runs Start a run on a project
GET /agents/runs List runs, filtered by projectId, status or parentRunId and paginated with page and pageSize
GET /agents/runs/{id} Get a run
GET /agents/runs/{id}/events List the run's events after the since sequence number
POST /agents/runs/{id}/cancel Cancel a run
POST /agents/runs/{id}/steer Send a message to the agent, optionally anchored to a line of its diff
POST /agents/runs/{id}/respond Approve or reject the agent's plan, or answer its question
POST /agents/runs/{id}/approve Start a run the preflight gate holds in awaiting_approval
POST /agents/runs/{id}/continue Start a new run that continues an ended run on the same branch and pull request

To comment a line of the agent's diff, add an anchor to the steering message. side is new for a line of the file as written or old for a removed line, and changeSeq is the sequence number of the file_change event whose diff you comment on:

curl -X POST \
  -H "X-API-Key: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Reuse the existing retry helper here",
    "anchor": {"path": "internal/billing/client.go", "line": 42, "side": "new", "changeSeq": 17}
  }' \
  https://api.archyl.com/api/v1/agents/runs/{runId}/steer

Schedules

Method Path Description
GET /agents/schedules List schedules, optionally filtered by projectId
POST /agents/schedules Create a schedule (5-field cron expression, evaluated in UTC)
PUT /agents/schedules/{id} Update a schedule
POST /agents/schedules/{id}/toggle Enable or disable a schedule
POST /agents/schedules/{id}/trigger Start a run from the schedule now
DELETE /agents/schedules/{id} Delete a schedule

MCP Connectors

Method Path Description
GET /agents/connectors List connectors
POST /agents/connectors Create a connector
PUT /agents/connectors/{id} Update a connector
POST /agents/connectors/{id}/toggle Enable or disable a connector
DELETE /agents/connectors/{id} Delete a connector
POST /agents/connectors/test Test a connection and list the server's tools

On top of the codes listed under Error Handling, these endpoints return 409 when a run is not in a state that allows the action or the preflight gate refuses it, 422 when your organization's AI provider or model cannot run managed agents, and 429 when no concurrency seat is free. Request and response schemas are in the OpenAPI reference.

Error Handling

API errors return standard HTTP status codes:

Code Description
400 Bad Request - Invalid parameters
401 Unauthorized - Invalid or missing API key
403 Forbidden - Insufficient permissions
404 Not Found - Resource doesn't exist
500 Internal Server Error

Error responses include details:

{
  "error": true,
  "message": "instructions are required to continue a run"
}

SDKs & Libraries

Coming soon:

  • JavaScript/TypeScript SDK
  • Python SDK
  • Go SDK

Use Cases

CI/CD Integration

Automatically update architecture after deployments:

- name: Update Architecture
  run: |
    curl -X POST \
      -H "X-API-Key: ${{ secrets.ARCHYL_API_KEY }}" \
      https://api.archyl.com/api/v1/projects/$PROJECT_ID/discover

Custom Tooling

Build internal tools that interact with your architecture:

  • Architecture validation
  • Compliance checking
  • Documentation generation

AI Assistants

Use MCP to let AI assistants understand and update your architecture:

  • Ask questions about your architecture
  • Create elements through natural language
  • Generate documentation automatically

API Documentation

The full interactive API documentation is available at:

https://api.archyl.com/docs

This OpenAPI documentation includes:

  • All available endpoints
  • Request/response schemas
  • Try-it-out functionality
  • Authentication examples

Next Steps